I started to reply to this topic almost a month ago, but somehow my reply got lost instead of posted. When I asked my colleagues that develop the calibration algorithm for the device, the best guess I got back was that the screen plane is not correctly set. What are the resulting screen plane coordinates when you use the setdisplayarea tool?
A couple of thoughts from looking at the video:
– The calibration points/targets seem quite large on the display – they are not too large to be able to focus your eye-gaze to the center of the circle?
– Would you say that looking at the four calibration points corresponds to looking at the corner of a 27″ or less display at arms length distance?
– Is the screen plane set to the full 46″ but the calibration points only to a smaller area of the screen?
– Is it individual points that are failing the calibration or the calibration as a whole that does not get enough data? (You can check the data per point by using the Get Calibration Point Data Items function).